Its the Metadata the matters...

Scott Karp has had another go at the value of data going forward, this time commenting on an earlier piece by Umair Haque, arguing data is a commodity.

Essentially Scott's argument is that data that is enclosed is not a commodity:

The data that is not a commodity (yet) is data that is NOT freely available on the web, e.g. the personal data we put on walled-garden social networking sites like Facebook.

It’s data about our lives that we choose to share only with friends and family, not with the whole world. It’s personal identification data, like birthday, address, and phone number, which we don’t want to share on the open web. It’s our searching habits, our purchase habits, our surfing habits, everything we do online.

Whereas Umair argues that, in essence, the data source adds value:

Think about it this way: the lower the cost of interaction, by definition, the more abundant data is - because every interaction creates reams of data. More data is created tomorrow than was created yesterday. And so on.

What is valuable are the things that create data: markets, networks, and communities.

I would argue that, although the value of rare data is not a commodity, and data creation entities are no doubt valuable, they are not actually the core area of value creation.

I'd go back to our earlier response to Scott's first post, and point out the real value is in the metadata - the data about the data. The core value creation stage is data Aggregation - seeking, structuring, sifting - and even sending & storing, rather than data Creation or data Distribution.

Hey - is that a 5-S model for data value

In their own ways, Search Engines (seek), Social Networks (structure), Tagging and Memetic tools (sift) etc are all adding meaning to the superabundance of data.

In my 20-odd years in consulting around IT, on the net and Web, in running businesses, and in managing operations I have never seen data do anything useful in and of itself (except if it is rare and ownership confers advantage). But once the relevant data is found, once it has been structured and sifted, once it is in the right hands, it is extremely powerful.

In other words, it becomes Information.

This may seem like hair-splitting, but having seen how many operations / processes / businesses gaily lose, or fail to capture, the metadata it is worth making this point again and again. In video, for example, the data about the media keeps on being scrubbed as it moves along its processing, making the problem of video search (and the need for end-of-process tagging) inefficient and expensive.

To re-iterate in terms of the Scott / Umair polarity - data in and of itself is (relatively) valueless, and if data producers are producing commodity data, then they too are relatively valueless. What is valuable is what was called (in the old days) Information, ie the ability to add structure and context to data.